Developing visual web-based log parsing and analysis tool for various kinds of log data

Abstract

Log analysis is utilized in many different software systems and has lots of uses cases from intrusion detection to analysing software bugs and other malfunctions. The anomaly detection has usage even in the automotive and aviation industries. The academic community is developing new ways to analyse logs in better ways, but available log data is still minimal, which is why the development of new methods is slower. The purpose of this research was to focus on developing ways on how to visualize the log analysis results. This type of visualization is helpful for software engineers to troubleshoot flaky software tests.

During the research, the objective was to develop a web-based prototype for log analysis. The research was carried out with a design science research method where the purpose was to evaluate the prototype with test subjects. The developed artefact demonstrated satisfactory performance with smaller log files. However, further adjustments were required when larger log files were inspected. The iterations for the developed web-based log analyzer were identified from evaluations and recommendations for future artefact are given. Overall, visualizing large amounts of data on web graphs remains a challenging task. Data reduction and other types of processing represent potential solutions to these types of issues.